Subject and citation indexing. Part I: The clustering structure of composite representations in the Cystic Fibrosis Document Collection

The presence of clustering structure in the cystic fibrosis (CF) Document Collection is evaluated as a function of the exhaustivity of five composite representations. The composite representations are constructed from two subject descriptions, based on Medical Subject Headings and subheadings, and two citation indexes, based on the complete set of references in and a comprehensive set of citations to each document. Experimental results reveal observable evidence for clustering structure at all exhaustivity levels of all composite representations but also show that the evidence for clustering structure diminishes as the exhaustivity of each representation is decreased. The representation composed of references and citations shows less evidence of clustering structure at the exhaustive level but more uniform evidence of clustering structure over a wide range of exhaustivity levels than composite representations that include subject descriptions. The structures imposed on the CF Document Collection by all composite representations satisfy the necessary condition for a meaningful clustering outcome. © 1991 John Wiley & Sons, Inc.